Crontab

• lxb1420:

# Prepares files for monitor (check the dat directory etc...)
34 * * * * /storage/lqcd/apps/output/mon/plot.py >> /storage/lqcd/apps/output/mon/plot.log 2>&1
# Prepares files for monitor (more preparation...)
35 * * * * /storage/lqcd/apps/output/mon/preparePlot.py >> /storage/lqcd/apps/output/mon/preparePlot.log 2>&1
# Keep the bak directory clean...
04 * * * * rm /storage/lqcd/apps/output/bak/snap*bak >> /storage/lqcd/apps/output/mon/rmbak.log 2>&1
24 * * * * rm /storage/lqcd/apps/output/bak/fort*bak >> /storage/lqcd/apps/output/mon/rmbak.log 2>&1
# Prepares the tar file (fort.15). It fails frequently...
30 7 * * * /storage/lqcd/apps/output/backup.sh >> /storage/lqcd/apps/output/mon/backup.log 2>&1

• lxb7232:

34 * * * * /storage/lqcd/apps/output/mon/plot.py >> /storage/lqcd/apps/output/mon/plot.log 2>&1
35 * * * * /storage/lqcd/apps/output/mon/preparePlot.py >> /storage/lqcd/apps/output/mon/preparePlot.log 2>&1
04 * * * * rm /storage/lqcd/apps/output/bak/snap*bak >> /storage/lqcd/apps/output/mon/rmbak.log 2>&1
24 * * * * rm /storage/lqcd/apps/output/bak/fort*bak >> /storage/lqcd/apps/output/mon/rmbak.log 2>&1
30 7 * * * /storage/lqcd/apps/output/backup.sh >> /storage/lqcd/apps/output/mon/backup.log 2>&1

• lxplus (acrontab):

# Count active workers for the 2 masters
34 * * * * lxplus /afs/cern.ch/user/l/laman/qcd/plot/lxb1420/active_workers >> /afs/cern.ch/user/l/laman/qcd/plot/lxb1420/active_workers.dat 2> /dev/null
33 * * * * lxplus /afs/cern.ch/user/l/laman/qcd/plot/lxb7232/active_workers >> /afs/cern.ch/user/l/laman/qcd/plot/lxb7232/active_workers.dat 2> /dev/null
# Copy relevant monitor fiels for plotting (should be done for both, now lxb7232 only due authenitcation problems
40 * * * * lxplus scp lxb7232:/storage/lqcd/apps/output/mon/fileRate.dat  /afs/cern.ch/user/l/laman/qcd/plot/lxb7232 >> /afs/cern.ch/user/l/laman/qcd/plot/lxb7232/trf.log 2>&1
40 * * * * lxplus scp lxb7232:/storage/lqcd/apps/output/mon/betaRate.dat  /afs/cern.ch/user/l/laman/qcd/plot/lxb7232 >> /afs/cern.ch/user/l/laman/qcd/plot/lxb7232/trf.log 2>&1
40 * * * * lxplus scp lxb7232:/storage/lqcd/apps/output/mon/histoRate.dat /afs/cern.ch/user/l/laman/qcd/plot/lxb7232 >> /afs/cern.ch/user/l/laman/qcd/plot/lxb7232/trf.log 2>&1
# Publish plots (basically runs gnuplot, not available on the master machines and copy results in public)
42 * * * * lxplus /afs/cern.ch/user/l/laman/qcd/plot/publishPlots.py >> /afs/cern.ch/user/l/laman/qcd/plot/publishPlots.log 2>&1
# Publish tar (fort.15). Again, for now only for lxb7232
00 8 * * * lxplus scp lxb7232:/storage/lqcd/apps/output/lxb7232.tgz  /afs/cern.ch/user/l/laman/public/lxb7232.tgz >> /afs/cern.ch/user/l/laman/qcd/plot/lxb7232/trf.log 2>&1

• Other useful things...
  • Under /storage/lqcd/apps/output the script "massimo" copies things to AFS (useful for lxb1420) and does some monitor (using log files). Note that the #files hours has to be multiplied by 3 in general (I count the number of results returned to the masters, usually corresponding to 3 iterations

-- MassimoLamanna - 23 Jun 2008